dtype kwargs of aggregation methods not passed to numpy function

[loicdtx]

Hi,

I'm trying to pass dtype= as kwargs to some aggregation methods, but it seems that the argument is ignored. See the example below.

import xarray as xr
import numpy as np
from datetime import datetime
import datetime as dt

arr = np.random.randint(0, 10000, size=(20, 1000, 1000), dtype=np.int16)
date_list = [datetime(2018, 1, 1) + dt.timedelta(delta) for delta in range(20)]
xarr = xr.DataArray(arr, dims=['time', 'x', 'y'], coords={'time': date_list})
xset = xr.Dataset({'blue': xarr, 'green': xarr, 'red': xarr})
print(xset)

# <xarray.Dataset>
# Dimensions:  (time: 20, x: 1000, y: 1000)
# Coordinates:
#   * time     (time) datetime64[ns] 2018-01-01 2018-01-02 2018-01-03 ...
# Dimensions without coordinates: x, y
# Data variables:
#     blue     (time, x, y) int16 1946 9194 1563 8318 938 8498 7999 6609 6275 ...
#     red      (time, x, y) int16 1946 9194 1563 8318 938 8498 7999 6609 6275 ...
#     green    (time, x, y) int16 1946 9194 1563 8318 938 8498 7999 6609 6275 ...

xset_mean = xset.mean('time', keep_attrs=True, dtype=np.int16)
print(xset_mean)

# <xarray.Dataset>
# Dimensions:  (x: 1000, y: 1000)
# Dimensions without coordinates: x, y
# Data variables:
#     blue     (x, y) float64 5.46e+03 4.717e+03 5.959e+03 4.31e+03 4.865e+03 ...
#     red      (x, y) float64 5.46e+03 4.717e+03 5.959e+03 4.31e+03 4.865e+03 ...
#     green    (x, y) float64 5.46e+03 4.717e+03 5.959e+03 4.31e+03 4.865e+03 ...

# Pure numpy
arr_mean = arr.mean(axis=0, dtype=np.int16)
print(arr_mean.dtype)
# int16

INSTALLED VERSIONS

commit: None
python: 3.5.2.final.0
python-bits: 64
OS: Linux
OS-release: 4.4.0-104-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: en_US.UTF-8
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

xarray: 0.10.0
pandas: 0.22.0
numpy: 1.14.0
scipy: None
netCDF4: 1.3.1
h5netcdf: None
Nio: None
bottleneck: None
cyordereddict: None
dask: 0.16.0
matplotlib: None
cartopy: None
seaborn: None
setuptools: 38.4.0
pip: 9.0.1
conda: None
pytest: None
IPython: None
sphinx: None

[fujiisoup]

Thanks for the report.
This bug was also reported in #1838, and was fixed by #1841.
I think the current master handles dtype argument for these aggregation methods.

[loicdtx]

Thanks @fujiisoup, indeed just saw #1838. Sorry for the duplicate 😬

[loicdtx]

And I confirm that installing xarray from HEAD of the master branch fixes the issue described. Thanks